My Brocade SLKWRM 300 is monitored by Nagios which has been reporting warning and critical states for all 3 of my brocade fans.
When I check my brocade fan status with fanshow, they all report as OK however I believe the fan speed is high and they are constantly fluctuating:
brocade:root> fanshow Fan 1 is Ok, speed is 8035 RPM Fan 2 is Ok, speed is 8035 RPM Fan 3 is Ok, speed is 7848 RPM brocade:root> fanshow Fan 1 is Ok, speed is 8132 RPM Fan 2 is Ok, speed is 7941 RPM Fan 3 is Ok, speed is 7848 RPM brocade:root> fanshow Fan 1 is Ok, speed is 8132 RPM Fan 2 is Ok, speed is 8035 RPM Fan 3 is Ok, speed is 7941 RPM brocade:root> fanshow Fan 1 is Ok, speed is 8132 RPM Fan 2 is Ok, speed is 8035 RPM Fan 3 is Ok, speed is 8132 RPM
What is the "normal" fan speed? What RPM is acceptable and at what speed do the fans start to experience an issue?
Please run tempshow and sensorshow commands from admin login on the switch and paste the findings here.
The output of commands tempshow and sensorshow is:
brocade:root> tempshowSensor Sensor State Centigrade Fahrenheit ID Index========================================================== 1 0 Ok 22 71 2 1 Ok 22 71 3 2 Ok 24 75
brocade:root> sensorshowsensor 1: (Temperature) is Ok, value is 22 Csensor 2: (Temperature) is Ok, value is 22 Csensor 3: (Temperature) is Ok, value is 24 Csensor 4: (Fan ) is Ok,speed is 8035 RPMsensor 5: (Fan ) is Ok,speed is 7941 RPMsensor 6: (Fan ) is Ok,speed is 8035 RPMsensor 7: (Power Supply) is Ok
Alrighty then. I presume you are asking about the nominal or normalized value of the fan speed because it seems higher than before, or than you are used to? Frankly, it seems higher than nominal to me, given the temps of the sensors involved. So let's see what we can tell from this info.
The swich goes through a POST process when first powered on that will run through the I2C bus and stress each element so that full capability is offfered during normal ops. You can hear this typically when a switch or dir is powered as the fans will start up, go to high state, then after 5-7 seconds will go to a lower or idling state as the sensors monitor and take control of the fan controller speed output.
In your case, it appears that given the temps of ALL the sensors in the chassis, that the fan controller did not take control of ALL the speed output, and the fans are all stuck on high speed. Since this is not related to one sensor, or one fan, but all together, it's more likely that the fan controller didn't complete it's POST process fully.
The semi-bad news is that there is no user access to the fan controller via the CLI unless you want to go in and start hacking at the code for fan controller. Very, very, not recommended. Plus, even if you did hack at the code that controls the fans, you would have to run another POST to get the deamon to start taking control of the fans speed again.
Things to check first: 1) Is this switch at/near the top of rack ,and is there limited cool air flow to the switch? If so, then it's possible that the amount and temp of the hot air avail for cooling is only available with the fans on high. Provide more, and cooler air to the chassis and see if that lowers the fan speed. Give it a few minutes with more cooling flow to adapt.
2) If you can set a maint window, the best way to test this would be to power off, and unplug the power supplies and start over. You will need to actually unplug the power cords, and let the volitile memory of the chassis lose it's current state before re-powering the unit. So - turn power off, remove power plug(s) from the unit, wait 2 min, plug in, turn on and allow the switch to power up.
If both of these do not reduce the fan speed, you may need to contact your maint vendor for further investigation.
Best of luck.
The brocade is near the top of the cabinet and there is a network switch directly racked above it where that network switch has the fan pointing to the front of the rack and I'm thinking it's blowing that hot air out and then down directly below to the brocade switch. I tried to block that direction as much as I could but it's still an issue.
The last time I rebooted the brocade was 6-8 months ago for a firmware upgrade. Even with that said, the fans have been fine only until the past week or two when they've been running high. The issue as well is that it is close to 100 degrees outside and as much as we try to keep that data center cool, it's never cool enough.
I plan to update the firmware patch in a month or so so the brocade will be rebooted at that time. I will try your suggestion of unplugging power cables and plugging back in . Hopefully that will resolve it.
Thank you for all your assistance!
Ya know what? I gotta call in artillery on my own position. I didn't read the last two words of your first thread, or didn't process them right. If the fans are fluctuating, then the fan speed controller is doing it's job, and as the sensors go down below a pre-set threshold the controller drops the fan speed to lower power usage. So - my bad, the fan controllers are doing their job, the fans are doing their job the Brocade switch is operating nominally - but you're blowing hot air from the rack into the switch and it can only keep cool on high speed(most of the time).
So - forget about rebooting the switch(a firmware download doesn't really count). Get it some cooler air, use a duct or move some stuff around in the cab and the world will once again be spinning in greased grooves.